Which FT3, FT4 and TSH levels have the highest and lowest prevalence rates for 10 common health disorders?
- Hyperlipidemia (high cholesterol)
- Coronary artery disease
- Heart failure
- Atrial fibrillation
- Peripheral vascular disease
- Renal failure (kidney failure)
Is high-normal TSH and low FT3 associated with one condition, while high-normal FT4 and low or high TSH strongly associated with another condition?
Or do they all have a generally similar pattern of strong disease associations with certain thyroid hormone levels?
Anderson et al, 2020 collected data from the medical charts of 174,914 adults from 1999 to 2018 and associated their diagnoses at the time of their FT4, TSH, and FT3 thyroid laboratory tests.
The researchers provided this rich treasure trove of data in appendix tables. However, their article focused only on atrial fibrillation risk and singled out the FT4 hormone because of their chosen data analysis methods.
Here I bring Anderson and team’s appendix data set into focus through visual and verbal analyses their article had no time or space to provide.
In the 10 data tables, the visual patterns will jump out through heat maps that color-code higher and lower prevalence rates in a grid of 18 hormone categories (6 levels per hormone x 3 hormones).
Each disease’s table comes with a scientific discussion of its puzzles and patterns that you can click to expand. Read quotations from other scientists that noticed similar patterns and commented on them.
At the end, tables summarize rankings of prevalence rates: One gives the average rank of prevalence rates across all 10 disorders and the other focuses on 4 cardiovascular disorders.
What can we learn from this comparison?
- Normal TSH levels rarely associated with the lowest rates of these disorders, which is unexpected given prevailing beliefs about the safety of normal TSH.
- Abnormally high or low TSH categories rarely had the highest prevalence rates for these disorders, which is also unexpected for the same reason.
- Instead, FT4 and FT3 levels usually had the highest and lowest prevalence rates.
- TSH and FT4 levels in high-normal range often had higher prevalence rates than extremely high or low levels, which is puzzling.
- FT3 and FT4 prevalence rates often trended in the opposite direction from each other. This shows that the human body responds differently to these two forms of thyroid hormone in circulation.
- Hormone categories with lower average FT3:FT4 ratios often correlated with higher disease prevalence rates. Therefore, the hormone relationships matter, not just their levels.
Why has it been so difficult for scientists and doctors to see these patterns emerge in disease prevalence rates? What are the implications of long term medical blindness to the power of FT3 levels and FT3:FT4 ratios in human health? What do the data mean for treated thyroid patients? I’ll give some suggestions for action.
Copyright fair dealing note
How did Anderson and team obtain their data?
In a health care region in the northwestern United States, medical records between 1999 and 2018 were searched.
- Adults over 18 years old, 174,914 were chosen if they had a FT4 lab test and were not treated with thyroid medications. The date of this FT4 test was considered their “study entry” date.
- Among these people, 147,834 had a TSH test within 60 days of their study entry.
- Among these people, 26,524 also had a FT3 test within 60 days.
On average, people had one, two, or three of these these tests within a period of only 2 days.
The population’s official diagnostic codes for atrial fibrillation and 9 other chronic diseases at the time of this set of tests were recorded. Therefore, the research team tabulated disease association rates with each subcategory of hormone levels.
Anderson’s article, titled “Free thyroxine within the normal reference range predicts
risk of atrial fibrillation,” focused on one disorder — atrial fibrillation (AF).
Why are these unadjusted prevalence rates worthy of analysis?
For advanced scientific readers: Click to expand
How to read Anderson’s data
In the tables below, each hormone has six levels from low to high:
- Normal Q1,
- Normal Q2,
- Normal Q3,
- Normal Q4,
The four divisions within in reference range are “quartiles.”
The reference ranges were:
- FT4: 0.75 to 1.50 ng/dL
- TSH: 0.54 to 6.80 mcIU/L (Units equal to mU/L; 0.4 to 4.0 is more common.)
- FT3: 2.40 to 4.20 pg/mL (100x smaller than the usual unit for FT3, pg/dL)
Where do the 4 Quartiles fit within reference range?
Anderson and team put all the people with “Normal FT4” and divided them into four equal sized groups, and did the same for the other two data sets. Dividing data into (3) tertiles, (4) quartiles or (5) quintiles is standard practice.
This resulted in each group covering a different portion of each reference range.
- Notice the pink gridlines representing the reference ranges for the hormones.
- Notice the size of Q4 on the right hand side of the diagram. Anderson’s “Normal Q4” quartile encompassed most or all of the upper half of reference range for each hormone. Q1, Q2, and Q3 cover narrower regions in the lower half of the reference ranges.
The TSH upper limit is controversial. In the region of the US where the study was conducted, the TSH limit was raised significantly near the very end of the study period. Their “Normal Q4” includes TSH values up to 6.80 mIU/L. (See “Details on methods” below). This means that some people with subclinical hypothyroidism were misclassified in Normal Q4.
Caution: These data are only from people on NO thyroid therapy.
The following prevalence rates are associations found in UNtreated individuals.
Therefore, these associations cannot translate to simple prescriptions or risk estimates for treated thyroid patients’ hormone levels.
Thyroid disease and its treatments distort TSH-FT4-FT3 hormone relationships even when they may “normalize” TSH and FT4.
The prevalence rates you see below will be different in various populations of treated thyroid patients on different thyroid hormone treatments — levothyroxine (LT4), liothyronine (LT3), porcine desiccated thyroid, and combinations.
There is no simple way to minimize disease risk in a hormone-treated person whose FT3:FT4 ratio and/or TSH-FT3 relationship is abnormal during thyroid therapy. In some patients with central hypothyroidism, the HPT axis is untrustworthy, and TSH is lower than it should be per unit of FT4 and FT3 because of dysfunctional TSH secretion.
Application of research must be approached with great caution to people who were not the main research population in this study. There is a little bit of data on treated patients’ atrial fibrillation risk in Anderson’s study, but their thyroid disease etiologies, antibodies, thyroid gland health, and dose changes were not followed.
The 10 chronic disorders
Hypertension prevalence rates: Discussion
Hyperlipidemia prevalence rates: Discussion
Depression prevalence rates: Discussion
Diabetes prevalence rates: Discussion
5. Coronary artery disease (CAD)
CAD prevalence rates: Discussion
6. Heart failure
Heart failure prevalence rates: Discussion
7. Atrial fibrillation
Atrial fibrillation (AF) prevalence rates: Discussion
8. Peripheral vascular disease (PVD)
PVD prevalence rates: Discussion
9. Renal failure
Renal failure prevalence rates: Discussion
Dementia prevalence rates: Discussion
For each of the 10 disorders, the 18 hormone level categories were ranked from lowest (1) to highest per set, using repeated numbers (5,5,5) when the prevalence rates were identical in two hormone level categories. This was the result:
The boldest colors are in the bottom row. FT3 spans an average prevalence rate ranking from 13.8 to 2.2. The difference between Low FT3 and Normal Q1 FT3 is almost double.
Beyond FT3, the secondary hot spot is TSH in Normal Q4, and the secondary cold spot is High FT4.
Narrowing down the ranking summary to focus only on 4 chronic cardiovascular diseases yields a similar pattern:
FT3 is still the most intense row. Now the step from Low FT3 to Normal Q1 is 3x the distance in the ranking of prevalence rates. High-normal TSH is still the next hot spot.
Now the red color is more evenly distributed. Notice that the 3 types of hypothyroidism (low FT4, high TSH, low FT3) are color coded red.
Why do doctors associate hyper with cardiovascular symptoms and tend to ignore the association with hypo?
These 10 chronic diseases cost our health care systems a lot of money and can cause great human suffering over a lifetime.
We can’t afford to separate thyroid hormone levels from chronic disease when they give us such major clues about how to manage these diseases more wisely.
Here are some suggested action items:
- Distinguish acute from chronic Low T3 syndrome (NTIS)
- Treat and prevent low T3 syndrome.
- Stop excluding thyroid patients from NTIS study and treatment.
- Call for a T3 paradigm shift in thyroid science
1. Distinguish acute from chronic Low T3 syndrome (NTIS)
The data say nothing about the duration of low T3 syndrome, but 15% of the FT3-tested population fit in that category.
The acute (and often benign) phase of NTIS must be distinguished from the pathological state of chronic low T3 that demonstrates a critical failure to recover a vital hormone supply (Moura Neto & Zantut-Wittmann, 2016; Van den Berghe, 2014).
Chronic low Free or Total T3 and chronic low T3:T4 ratios are not just effects of illness, but can also become pathological drivers and maintainers of illness.
Many decades of research in NTIS have shown that a weak TSH response and/or thyroid gland response can perpetuate a low Total or Free T3 level and T3:T4 ratio, leading to increased rates of poor health outcomes and death. (NOTE: In severe illness, the Total T3 is often used instead of Free T3 because some illnesses and blood thinners like heparin can artificially inflate Free hormone test results.)
Look at the T3:FT4 ratios of people with heart failure:
“Lower levels of total T3 were well correlated with 1-year HF in PCI-treated AMI patients. The T3/fT4 levels can be an additional marker to predict HF.”Kang et al, 2017
The chronic low T3 levels and T3:T4 ratios during chronic disease are prognostic of future problems.
This is the case in many types of chronic illness including renal failure, cardiovascular diseases, not just in acute conditions demanding intensive care (Ataoglu et al, 2018). (See our review of Ataoglu: Low T3 in critical illness is deadly, and adding high T4 is worse.)
2. Treat and prevent chronic low T3 syndrome.
Enough research has catalogued the high rates of death and morbidity in Low T3 Syndrome (also called mistakenly called “euthyroid sick syndrome” just because TSH usually remains normal).
Yet publications continue to repeat the skeptical medical stasis that
“no clear consensus has emanated from clinical studies”
of thyroid hormone interventions, so
“many scholars suggest that treatment of the primary disease takes precedence over ESS [euthyroid sick syndrome] intervention.”Wang et al, 2018
However, quotes like these reveal an arbitrary, false division between the “primary disease” (such as heart failure) and a thyroid hormone metabolism disorder such as ESS, when thyroid hormone signaling clearly modifies primary disease outcomes.
Thyroid scientists with courage have previously advocated for more NTIS treatment trials involving T3 (DeGroot, 2006), but it seems to have fallen on deaf ears. The considerations of treatment are complex, but few have the courage to face them for the sake of saving lives, health care dollars, and human suffering.
If we supply ventilators and oxygen to people who cannot breathe, then it is unethical not to assist the thyroid-healthy human body to recover healthy FT3 levels and FT3:FT4 ratios to prevent the progression of a chronic disease.
Patients with chronic diseases deserve to access FT3 thyroid hormone levels and ratios that are conducive to their recovery.
3. Stop excluding thyroid patients from NTIS study and treatment
This research by Anderson and team unfortunately excluded from their main data set all people dosed with thyroid hormone because of our complexity.
There were 37,288 people excluded from Anderson’s study for this reason. No data tables were provided for them, no explanation of their quartile divisions, and no FT3 or TSH prevalence data was provided. We were only given a verbal summary of the FT4 odds ratios that the researchers felt were most important as they looked for linear trends within the reference range.
This is a significant loss of data. We have insufficient knowledge of a neglected population.
It is very disturbing that thyroid patients are routinely excluded from low T3 syndrome / nonthyroidal illness (NTIS) research, as we’ve discused in previous posts.